# Verb-initial grammars: A multilingual/parallel perspective
# ESRC Project RES-000-23-0505
# Oxford University
# Charles Randriamasimanana


# Malagasy syntax/semantics: 
# Malagasy Pronouns - ESRC-OX-05-CR206.

# In addition to the basic process outlined and illustrated in Randriamasimanana
# (1986: 326-386), the following additional information may be relevant pertaining to
# the distribution of different pronoun forms in this language. Towards the end of
# this document, some illustrative examples involving reflexives will be offered.


# Testfile:Malagasy.pronoun.syntax.txt

# File 1: Testfile for pronouns.

# Malagasy has three types of pronoun forms: (A) independent pronoun forms (B) empty
# positions of which zero pronoun is one instance and (C) clitic forms attached to a head, 
# i.e. a V, an N or a Prep.

# A. Independent pronoun forms:
# Independent pronoun forms come in two varieties: simple forms and complex ones;

# Simple forms include:
# izaho 'I.nominative.emphatic', aho 'I.nominative.nonemphatic';
# ahy 'me.accusative';
# isika 'we.inclusive.nominative', antsika 'us.inclusive.accusative';
# izahay 'we.exclusive.nominative', anay 'us.exclusive.accusative';
# ianao 'you.sing.nominative', anao 'you.sing.accusative';
# ianareo 'you.plural.nominative', anareo 'you.plural.accusative';
# ise 'you.nonDistant/Intimate.sing.nominative', an'ise 'you.nonDistant/Intimate.sing.acc.'
# izy '(s)he.sing.human.nominative', azy 'him/her.sing.human.accusative';

# Interrogative pronouns iza 'who.human.nominative', an'iza 'whom.human.accusative';
# there is an Unmarked use as well as a Marked one, i.e. via extension & feature erasure. 
# See examples below.

# Complex forms can take one of two shapes:

# Shape A: pronoun izy/azy + plural DX like 'plural.demonstrative.ireo' to refer to a human entity, 
# especially when an indication of plurality is required within the clause;
# izy ireo'(s)he.sing.human. + DX.plural', azy ireo'him/her.sing.human.accusative + DX.plural';
# also note that the DX will indicate plurality; in certain cases like 'izy mianaka''they.parents&
# chilren', 'izy mivady''they.spouses', it is also possible to have an appositive NP replacing DX 
# to make more precise the nature of the relationship between the entities being referred to.

# In addition to the above feature, the selection of the appropriate form of DX to use to refer
# back to a previously mentioned entity also depends on other factors which remain to be elucidated;
# for example, instead of the complex form 'izy ireo''plural.demonstrative.ireo', another form
# like 'ry zareo' D.plural + zareo''they' may be used; see illustration below.

# Shape B: ilay + izy to refer to a non-human entity, when the referent is singular; however when
# the referent is a plural, then a simple plural demonstrative like 'ireo' can be used;
# ilay izy 'it.sing.nonhuman.proDX.nominative + izy', an'ilay izy 'it.sing.nonhuman.proDX.accusative + izy';
# ireo'they.plur.nonhuman.proDX.nominative', an'ireo 'them.plur.nonhuman.proDX.accusative';
# where proDX is the pronoun counterpart of a DX type of determiner;


# Illustrative examples of simple pronoun forms:

  Handeha aho.
# H-andeha  aho.
# fut-go    I.nominative.nonemphatic
# 'I will go.'
# Note that the nonemphatic pronoun shows up after the VP;

# Ungrammatical
  Handeha izaho.(0! 0 0 0)
# H-andeha  izaho.
# fut-go    I.nominative.emphatic
# 'I will go.'
# This sequence is ruled out in standard Malagasy; however, it is possible
# in Child Speech especially in its truncated form 'zah;

  Izaho no handeha.
# Izaho                 no    h-andeha.
# I.nominative.emphatic focus fut-go
# 'It is I who will go.'
# Note that the emphatic first person pronoun is to the left of the focus particle no, i.e.
# basically to the left of VP;

# Ungrammatical
  Aho no handeha.(0! 0 0 0)
# aho                      no    h-andeha.
# I.nominative.nonemphatic focus will-go
# 'It is I who will go.'
# This sequence is irretrievably ungrammatical.

# Typically a grammatical subject is encoded as nominative; nevertheless in
# certain dialects (like Betsileo) and with certain verbs this may not always
# be the case
  Handeha anay. 
# H-andeha anay.
# fut-go   us.accusative
# "Us will go.'
# This use of the accusative case to mark a grammatical subject is dialectal.


# Unmarked use of interrogative IZ-particle 'iza''who.human.nominative':
  Ankizy iza no nandrovitra ity boky ity?
# Ankizy iza     no    n-an-d-rovitra               ity  boky ity?
# Child  IZ.part focus past-pref.an-ep.d-be.torn.up this book this
# 'Which child has torn this book up?'
# The above sequence is adapted from Rajemisa-Raolison (1995: 455).
# Interrogative particle IZ = WH, which shows up after the NP 'ankizy''child' unequivocally
# refers to a human.

# Unmarked use of interrogative pronoun 'iza''who.human.nominative':
  Iza no nanambara izany vaovao izany taminao?
# Iza           no    n-an-ambara            izany vaovao izany t-amin(a)-nao?
# Who.human.nom focus past-pref.an-be.stated that  news   that perf-to-clitic.you
# 'Who told you that piece of news?'
# The above sequence is adapted from Rajemisa-Raolison (1995: 455).
# Interrogative pronoun Iza = WHO unequivocally refers to a human.

# Extended use of interrogative pronoun 'iza''who.human.nominative' to non-humans, i.e.
# specifically to animates:
  Iza no hoentinao amin'ireto omby ireto.
# Iza           no    ho-entin(a)-(n)ao   ø-amin(a)           ireto omby ireto
# Who.human.nom focus fut-be.taken-by.you non-past-from.among these cows these
# 'Which one.singular is it will be taken by you, from among these cows?'
# The above sequence is adapted from Rajemisa-Raolison (1995: 455).
# Interrogative pronoun Iza can now refer to a non-human as there is an adjunct
# PP amin'ireto omby ireto, which makes explicit this specification, thus erasing
# the inherent feature 'human'.


# Extended use of interrogative pronoun 'iza''who.human.nominative' to non-humans, i.e.
# specifically to non-animates:
  Satroka iza amin'ireto no hovidinao?
# Satroka iza ø-amin(a)           ireto no    ho-vid(i)-i(na)-nao?
# Hat     IZ  non-past-from.among these focus pass.fut-root.vid(i)-suf.i(na)-by.you
# Literally 'Which one of these hats will be bought by you?''Which one of these hats
# will you buy?'
# The PP ø-amin'ireto 'non-past-from among these' is an adjunct to the verb 'hovidinao'
# which is in the future tense; the PP itself is in a non-past tense, which therefore can
# accommodate the future tense; as a result, we have the configuration 'future...future',
# the hallmark of adjunction in Malagasy. Now the use of the non-human plural DX ireto in
# this PP-adjunct erases the inherent feature 'human' characteristic of the IZ-particle.




# [ +/- Animate ] is a crucial feature of Malagasy pronouns.

# First, the case of [ +/- Human ]:
  Manasa lamba izy.
# M-an-(s)asa            lamba izy.
# pres-pref.an-root.sasa linen s/he.nom
# '(S)he washes linen.'
# Note that 'izy' can refer to a male or a female person; typically it refers to a human;
# however 'izy' cannot refer to a non-animate entity;

  Nahita azy i Soa.
# N-a-hita            azy         i    Soa.
# past-pref.a-be.seen him/her.acc D.sg Soa
# 'Soa saw him/her (human).'
# Note that 'azy' can refer to a male or a female person; typically it refers to a human;
# however 'azy' cannot refer to a non-animate entity;


# Illustrative examples of complex pronoun forms:

# First, the case of [ +/- Human ]:
  Manasa lamba izy mianaka.
# M-an-(s)asa       lamba izy      mianaka.
# pres-pref.an-sasa linen s/he.nom parents&children
# 'They, parents and children, wash linen.'

  Manasa lamba izy ireo.
# M-an-(s)asa            lamba izy      ireo
# pres-pref.an-root.sasa linen s/he.nom pluralDX
# 'They wash linen.'
# The complex nominative form 'izy ireo' unequivocally indicates that the referent is a plural
# and that it is human;

  Nahita azy ireo i Soa.
# N-a-hita            azy         ireo     i    Soa.
# past-pref.a-be.seen him/her.acc plurDX D.sg Soa
# 'Soa saw them (human).'
# The complex accusative form 'azy ireo' unequivocally indicates that the referent is a plural
# and that it is human;


# Second, the case of [ +/- Animate ]:
  Nahita an'ilay izy i Soa.
# N-a-hita            an'ilay izy   i    Soa.
# past-pref.a-be.seen acc.ilay.izy  D.sg Soa
# 'Soa saw it (sing.nonhuman).'
# This feature of Malagasy pronouns was already illustrated in Randriamasimanana
# (1986: 370-1, ex. (72)c as well as ex. (73)a, b and c; note that 'ilay izy' 'DX + izy',
# i.e. henceforth proDX + izy can only refer to a non-animate entity.


# Ungrammatical:
  Nahita ny mpanjaitra ny lamba aho.(0! 0 0 0)
# N-a-hita         ny         mp-an-jaitra           ny         lamba   aho.
# past-pref.a-hita Det.unspec pref.mp-pref.an-zaitra Det.unspec clothes I
# Intended reading: 'I saw the tailor of the clothes'(habitual reading)
# The above sequence is adapted from Dimitrios Ntelitheos. Malagasy Participant Nominalizations: 
# A Structural Account, AFLA XII, 31 May 2005 UCLA. 
# This sentence is anomalous; at best, it could be either Foreigner Talk or Child Speech:
# On the one hand, the headverb/noun 'mpanjaitra' means 'someone who habitually sews, as a
# profession'; on the other hand, the phrase 'ny lamba' refers to a 'specific set of clothes
# presumably already identified and established in a previous discourse'. The anomaly comes
# from the incongruous merging of the HABITUAL head 'mpanjaitra' & the non-HABITUAL complement
# 'ny lamba''the previously mentioned clothes'; this yields something like 'the tailors who
# keep sewing the same (previously mentioned) pieces of clothes over and over as a profession'.


# Ungrammatical:
  Nahita ny mpanjaitra azy aho.(0! 0 0 0)
# N-a-hita         ny         mp-an-jaitra            azy     aho.
# past-pref.a-hita Det.unspec pref.mp-pref.an-zaitra him/her I
# Intended reading:'I saw the tailor of it.'
# Actual reading: 'I saw the tailor of him/her.'
# The above sequence is adapted from Dimitrios Ntelitheos. Malagasy Participant Nominalizations: 
# A Structural Account, AFLA XII, 31 May 2005 UCLA. 
# The pronoun 'azy''him/her' is supposed to replace the phrase 'ny lamba''the clothes' in the
# previous sentence (immediately above).


# Ungrammatical
  Ny radara dia Rabe no nilaza fa Rasoa no nanao azy.(0! 0 0 0)
# Ny radara dia   Rabe no    n-ilaza  fa   Rasoa no    n-anao  azy.
# Det radar topic Rabe focus past-say that Rasoa focus past-do him/her?
#'As for the radar, it was Rabe who said that Rasoa built it.'
# This sequence is adapted from Ileana Paul 'Multiple Topics: Evidence from Malagasy'
# in Proceedings of AFLA-9, Cornell University Working Papers in Linguistics
# Volume 19, Spring 2003, ed. by Anastasia Riehl and Thess Savella.
# Note animate pronoun azy = him or her referring to a non-animate NP 'ny radara''the radar';

  Ny radara indray dia Rabe no nilaza fa Rasoa no nanao an'ilay izy.
# Ny radara indray  dia   Rabe no    n-ilaza  fa   Rasoa no    n-anao  an'ilay izy.
# Det radar  as.for topic Rabe focus past-say that Rasoa focus past-do acc.proDX.sg.nonhuman
#'As for the radar, it was Rabe who said that Rasoa built it.'
# Note the complex nonanimate form an'ilay izy = accusative + proDX.sg.nonhuman.acc.it;

  Nahita azy ireo i Soa.
# N-a-hita            azy        ireo          i    Soa.
# past-pref.a-be.seen him/her.acc.proDX.plural D.sg Soa
# 'Soa saw them (human beings).'
# Note complex animate form azy + ireo = azy + DX.plural, 'them-human.beings + DX.plural';

# Ungrammatical
  Ny vilia dia isan'andro Rabe no manasa azy ireo.(0! 0 0 0)
# Ny  vilia  dia   isan'andro Rabe no    m-anasa   azy ireo.
# Det dishes topic everyday   Rabe focus pres-wash them (humans)
#'As for the dishes, it's everyday that Rabe washes them.'
# Adapted from Ileana Paul Proceedings of AFLA-9, Vol. 19, Spring 2003,
# Cornell University
# Note the complex animate form azy ireo = them.human.being.accusative + DX.plural 
# referring to a non-animate NP 'ny vilia''the dishes';

  Ny vilia indray dia isan'andro Rabe no manasa an'ireo.
# Ny  vilia indray  dia   isan'andro Rabe no    m-anasa   an'ireo.
# Det dishes as.for topic everyday   Rabe focus pres-wash acc. proDX.plur.nonhuman
#'As for the dishes, it's everyday that Rabe washes them.'
# Note an'ireo = accusative + proDX.non-animate refers to a nonhuman NP 'ny vilia';
# with a proDX form like 'ireo' the accusative case-marker 'an' is mandatory;

# Ungrammatical
  Ny vilia indray dia isan'andro Rabe no manasa ireo.(0! 0 0 0)
# Ny  vilia indray  dia   isan'andro Rabe no    m-anasa   ireo.
# Det dishes as.for topic everyday   Rabe focus pres-wash proDX.plur.nonhuman
#'As for the dishes, it's everyday that Rabe washes them.'
# This last sequence is irretrievably ungrammatical since the accusative marker
# 'an' has been left out in front of the proDX form 'ireo'; this is different
# from the behaviour of DX 'ireo' as a determiner; see Document ESRC-OX-05-CR101/201;


# [ +/- Singular ] is a crucial feature of Malagasy pronouns.

  Nahita azy i Soa.
# N-a-hita            azy         i    Soa.
# past-pref.a-be.seen him/her.acc D.sg Soa
# 'Soa saw him/her (sing.human).'
# Azy is the singular form for the accusative case of a human pronoun;

  Nahita azy ireo i Soa.
# N-a-hita            azy        ireo          i    Soa.
# past-pref.a-be.seen him/her.acc.proDX.plural D.sg Soa
# 'Soa saw them (human beings).'
# Azy ireo = accusative.human.beings.pron + DX.plural, them-human.beings;
# To indicate plurality on the pronoun 'azy' a DX plural form like 'ireo'
# is added to it;

  Nahita an'ilay boky i Soa.
# N-a-hita            an'ilay     boky  i    Soa.
# past-pref.a-be.seen acc.DX.sing book   D.sg Soa
# 'Soa saw the (previous.mention) book (sing.nonhuman).'
# Note that DX ilay only exists in the singular form and is used to indicate
# that the accompanying pronoun is nonhuman; see immediately below;

  Nahita an'ilay izy i Soa.
# N-a-hita            an'ilay izy             i    Soa.
# past-pref.a-be.seen acc.proDX.sing.nonhuman D.sg Soa
# 'Soa saw it (sing.nonhuman).'
# Note the proDX form 'ilay izy' accompanied by the accusative case-marker an;
# unlike its DX cousin, proDX requires an overt case-marker; see evidence shown
# immediately below;

# Ungrammatical
  Nahita ilay izy i Soa.(0! 0 0 0)
# N-a-hita            ilay izy             i    Soa.
# past-pref.a-be.seen proDX.sing.nonhuman D.sg Soa
# 'Soa saw it (sing.nonhuman).'
# Note that proDX.sing.nonhuman 'ilay izy' requires an accusative case-marker;
# otherwise the sequence becomes irretrievably ungrammatical;

  Nahita an'ireo boky i Soa.
# N-a-hita            an'ireo     boky  i    Soa.
# past-pref.a-be.seen acc.DX.plur book   D.sg Soa
# 'Soa saw the (previous.mention) books (plural.nonhuman).'
# Note that ireo in 'ireo boky' is a DX.plur.nonhuman 'ireo' and that it can take
# an accusative case-marker although it does not have to because it is merely
# a determiner;


  Nahita ireo boky i Soa.
# N-a-hita            ireo     boky  i    Soa.
# past-pref.a-be.seen DX.plur book   D.sg Soa
# 'Soa saw the (previous.mention) books (plural.nonhuman).'
# Note that DX.plur.nonhuman 'ireo' does not require an accusative case-marker
# as would normally be the case with a pronoun, i.e. a proDX form;

  Nahita an'ireo i Soa.
# N-a-hita            an'ireo         i    Soa.
# past-pref.a-be.seen acc.proDX.plur  D.sg Soa
# 'Soa saw them (plural.nonhuman).'
# Note that proDX.sing.nonhuman 'an'ireo' requires an accusative case-marker;
# otherwise the sequence becomes irretrievably ungrammatical;

# Ungrammatical:
  Nahita ireo i Soa.(0! 0 0 0)
# N-a-hita            ireo         i    Soa.
# past-pref.a-be.seen proDX.plur  D.sg Soa
# 'Soa saw them (plural.nonhuman).'
# Note that proDX.plural 'ireo' requires an accusative case-marker;
# otherwise the sequence becomes irretrievably ungrammatical;

# Ungrammatical:
  Novorin'ny mpampianatra ny ankizy; nefa tsy nanana fotoana izy.(0! 0 0 0)
# No-vor(i)-in(a)  ny        mpampianatra ny        ankizy;  nefa tsy n-anana  fotoana izy.
# past-gathered-by Det.unspec teachers   Det.unspec children; but neg past-have time   s/he
# Intended reading:'The children were gathered by the teachers; but they did not have time.'
# The above sequence is adapted from Lisa Travis 'States, abilities, and accidents',
# 2005 Austronesian Formal Linguistics Association XII meeting at UCLA, 30 April.
# This sequence is ungrammatical for two main reasons: (i) a mismatch between
# the singular feature of 'izy' 's/he' and the plural feature of the putative antecedent 
# 'ny ankizy''the children'; (ii) the selection of the target form 'izy ireo', which is simply 
# not appropriate in this case. The pronoun 'izy' is a third person human singular form, 
# which is supposed to refer back to the definite phrase 'ny ankizy''the children' i.e. a plural.
# Presumably the other complex form 'ry zareo''they' would have been more appropriate,
# given the context of the discourse.

# Ungrammatical:
  Nivory ny olona; nefa tsy nanana fotoana izy.(0! 0 0 0)
# N-i-vory         ny         olona;  nefa tsy n-anana   fotoana izy.
# past-pref.i-vory Det.unspec people; but  neg past-have time    s/he
# Intended reading: 'The people met, but they did not have time.'
# The above sequence is adapted from Lisa Travis 'States, abilities, and accidents',
# 2005 Canadian Linguistics Association meeting, University of Western Ontario, 31 May.
# This sequence is ungrammatical for two main reasons: (i) a mismatch between
# the singular feature of 'izy' 's/he' and the plural feature of the putative antecedent 
# 'ny olona''the people'; (ii) the selection of the form 'izy ireo', which is simply 
# not appropriate in this case. See complex constructions below involving both
# coordinate and subordinate structures.
# The pronoun 'izy' is a third person human singular form, which is supposed to refer
# back to the definite phrase 'ny olona''the people' i.e. a plural.



# B. Empty pronoun positions:
# First, we will consider discourse-based null subject.

  Manao ahoana indray? 
# M-anao ahoana indray  [ ø ]? 
# pres-do how once.more zero.pronoun
# 'How are (you)?'
# This was an original sequence I received on Thursday 24 december 1998;
# this was the very first sentence in the e-mail message in question.
# Note the [ ø ]subject pronoun form, i.e. the personal pronoun ianao 'you.nominative.sing' 
# has been left out; note also that the verb does not encode any features such as person,
# gender, number and the like so that pro-drop following Romance languages like
# Italian or Spanish is not a very plausible analysis;


# By contrast, when the sequence contains deictics such as ireo...ireo 'these', etc
# surrounding a (non-)verbal predicate, the grammatical subject is mandatory; note
# that this is an open-ended process in that any predicate can be surrounded  by
# deictics such as ireo...ireo 'these', etc, as illustrated below.


# The case of verbal predicates:

  Ireo miady ireo ny zanakao.
# Ireo  m-iady     ireo  ny  zanaka-o.
# these pres-fight these Det children-of.yours
# 'Your children are there, fighting.'
# This sequence is from Abinal & Malzac (1888:282); note the plural deictics
# surrounding the VP miady; their presence makes the overt grammatical subject
# 'ny zanaka-o' 'the children-of-ours' obligatory; for further elaboration on this
# point, consult Randriamasimanana (2002:52-55); because the deictic encodes
# number, i.e. singular or plural nature of the grammatical subject of the clause
# the label AGR(eement) was used;

# Ungrammatical
  Ireo miady ireo.(0! 0 0 0)
# Ireo  m-i-ady         ireo   ø .
# these pres-pref.i-ady these zero.subject
# '.... are there, fighting.'
# This sequence where an overt grammatical subject is missing is irretrievably
# ungrammatical as an initial utterance in some discourse; contrast with the previous sentence.


# The case of non-verbal predicates (only a Prep predicate is illustrated below):

  Ireo ao ireo ny ankizy.
# Ireo  ø-ao               ireo  ny  ankizy.
# these nonperf-prep.there these Det child(ren)
# 'The children are there near and visible';
# note the plural feature of the deictic 'ireo' passed onto the subject NP 'ny ankizy;

# Ungrammatical
  Ireo ao ireo. (0! 0 0 0)
# Ireo  ø-ao               ireo  ø.
# these nonperf-prep.there these zero.subject
# '... are there near and visible';
# This sequence where an overt grammatical subject is missing is irretrievably
# ungrammatical as an initial utterance in some discourse; contrast with the previous sentence.

  Io ao io ny ankizy.
# Ireo  ø-ao               ireo  ny  ankizy.
# these nonperf-prep.there these Det child(en)
# 'The child is there near and visible';
# note the singular feature of the deictic 'io' passed onto the subject NP 'ny ankizy';


# The above remarks hold true of DX type of determiners with the possible exception of
# DX ilay 'the previous mention', ity 'this.singular.near.speaker'; this means that we
# we can have a DX Y DX configuration involving ity...ity, ireto...ireto, iny...iny
# ireny...ireny, iry...iry, irery ... irery, itsy ...itsy, iretsy... iretsy, iro...iro
# iro ... iro; and where Y may stand for VP, PP or AdjP.

# Recall that DX type of determiners contain at least the following items = 
# { ilay 'the previous mention', ity 'this.singular.near.speaker', ireto 
# 'these.plural.near.speaker', iny 'this.singular.near.hearer', ireny 'these.
#  plural.near.hearer', iry 'this.visible.far', irery 'these.plural.visible.far', 
# itsy 'this.singular.to.the.side', iretsy 'these.plural.to.the.side',
# iro ''this.singular.visible', irero 'these.plural.visible', ...}.



# Second, in the case of a complex construction, i.e. coordinate as well as subordinate structures, 
# especially those involving embedded Control, an empty subject pronoun is the norm; 

# Coordinate structures involving topic particle dia:

  Nitsangana izy dia niteny.
# N-i-tsangana         izy  dia   n-i-teny         [ ø ].
# past-pref.i-tsangana s/he topic past-pref.i-teny zero.pronoun
# '(S)he stood up and (subsequently) said.'
# This sequence is from Rajemisa-Raolison (1995: 167) and shows that even when a complex
# sentence is involved, i.e. we have two different verbs and two different sentences
# the second verb niteny 'past-speak' simply does not take an overt subject pronoun;

# Illegal use of topic-marker dia with an overt subject pronoun in a simple sentence:
# Ungrammatical
  Rasoa dia manasa lamba izy.(0! 0 0 0)
# Rasoa dia   m-an-(s)asa       lamba   izy.
# Rasoa topic pres-pref.an-sasa clothes s/he
# 'Rasoa, she's washing clothes.'
# This sequence was first proposed by Edward L. Keenan (1976:272, ex. 65) as of dubious
# grammaticality --indicated by ?. In fact, this sentence is simply not Malagasy. At best,
# it could be Foreigner Talk as it is not at all likely that the pronoun izy 's/he' would be 
# coreferential with the topic Rasoa inside the same sentence; another possibility would
# be Child Speech;

# Ungrammatical
  Rasoa omaly dia nanasa ny lamba izy.(0! 0 0 0)
# Rasoa omaly     dia   n-an-(s)asa       ny  lamba   izy.
# Rasoa yesterday topic past-pref.an-sasa Det clothes s/he
#‘Rasoa, yesterday, she washed the clothes.’
# This sequence is from Heather Flegg 'Topics and Clitic Left Dislocation in Malagasy'
# AFLA-10 28-30 March 2003. Austronesian Formal Linguistics Association, 10th. U HI-Manoa, 
# Honolulu, HI; the same remarks as above apply in this case as well.


# Coordinate structures involving coordinate conjunction ka:

  Nasainao aho ka tonga izao.
# N-asa-inao                        aho ka   tonga   [ ø ]        izao.
# past-root.asa-suffix.in(a)-by.you I   part arrived zero.pronoun now
# Literally 'I was invited by you and so (I) am arrived now.'
# 'You have invited me and as a result, here I am.'
# This sequence is from Abinal & Malzac (1888: 301) and shows that when particle ka
# is used to conjoin two different sentences the second verb, i.e. in this case, tonga 
# 'perf-arrive' simply does not take an overt subject pronoun;

  Nihazakazana izy ka lavo.
# N-i-hazakazaka         izy  ka   lavo                     [ ø ].
# past-pref.i-hazakazaka s/he part root.fallen.accidentally zero.pronoun
# Literally 'S/he was running and as a result, fell accidentally.'
# This sequence is from Rajemisa-Raolison (1995: 467) and shows when particle ka
# conjoins two different sentences, the second verb lavo 'fell.accidentally' simply
# does not take an overt subject pronoun;

# Ungrammatical:
  Nanomboka namaky ny taratasy ny mpianatra ka menatra izy.(0! 0 0 0)
# N-an-(t)omboka       n-a(n>m(-(v)aky   ny  taratasy ny  mpianatra ka   menatra izy.
# past-pref.an-tomboka past.pref.an-vaky Det letter   Det students  part ashamed s/he
# 'The student began to read the letter and he was embarrassed.'(Original translation)
# This sequence has been adapted from Polinsky and Potsdam 'Malagasy Control and Its 
# Theoretical Implications'. Also see Polinsky, Maria, and Eric Potsdam. 2002a. 
# Backward control: Evidence from Malagasy. In MIT Working Papers in Linguistics 44: 
# The Proceedings of AFLA XIII, 257-272.2002a.Backward control: Evidence from Malagasy.
# At best, this sequence is Foreigner Talk; in standard Malagasy 'ny mpianatra'
# and the pronoun 'izy' have to be disjoint in reference; furthermore the definite
# phrase 'ny mpianatra' would normally refer to a set of entities comprising several
# students, i.e. at the very least --especially in case of emphasis-- we should have 
# the plural form 'izy ireo'. At any rate, it is more natural in Malagasy to have
# a zero-pronoun, i.e. a pronominal form here when no emphasis is intended.



# Subordinate structures involving topic particle dia:

  Nony tonga tao izy dia nanao hoe.
# Nony tonga       t-ao       izy  dia   n-an-(t)ao        [ ø ]        hoe.
# when perf-arrive perf-there s/he topic past-pref.an-done zero.pronoun thus
# 'When s/he arrived there, s/he spoke thus.
# This sequence is from Abinal & Malzac (1888: 99) and shows that even when a complex
# sentence is involved, i.e. we have two different verbs and two different sentences
# the second verb, i.e. nanao 'past-do' simply does not take an overt subject pronoun;

  Nony tonga tao izy dia niteny hoe.
# Nony tonga   t-ao       izy  dia   n-i-teny         [ ø ]        hoe.
# when arrived perf-there s/he topic past-pref.i-teny zero.pronoun thus
# 'When s/he arrived there, she spoke thus.
# This sequence is from Rajemisa-Raolison (1995: 167) and shows that when particle ka
# is used to subordinate one sentence, the second verb, i.e. niteny 'past-speak' simply
# does not take an overt subject pronoun;

# Ungrammatical
  Raha manasa lamba Rakoto, dia omeko sakafo izy.(0! 0 0 0)
# Raha m-an-(s)asa lamba Rakoto, dia omeko sakafo izy.
# 'If Rakoto washes clothes, then I give him food.'
# This sequence has been adapted from Lisa Travis (in press);
# Note the illegal overt subject pronoun coreferential with Rakoto.
# Again, this is at best either Foreigner Talk or Child Speech;
# in standard Malagasy, this is simply ungrammatical unless
# izy and Rakoto are disjoint in reference.

# Ungrammatical
  Omeko sakafo Rakoto, raha manasa lamba izy.(0! 0 0 0)
# Ome-ko        sakafo Rakoto, raha m-an-(s)asa       lamba   izy.
# be.give-by.me food   Rakoto  if   pres-pref.an-sasa clothes s/he
# 'I give Rakoto food if he washes clothes.'
# This sequence has been adapted from Lisa Travis (in press);
# Again, this is at best either Foreigner Talk or Child Speech;
# in standard Malagasy, this is simply ungrammatical unless
# izy and Rakoto are disjoint in reference.



# Zero-pronoun and verbs of saying:
# Apparently the question of zero-pronoun simply does not arise with a verb of saying.

  Nolazain'i Paoly hovidina ilay boky.
# No-laza-in(a)             i    Paoly  ho-      vid(i)-  ina         
# past.pass-say-pass.ina.by D.sg Paul   fut.pass-root.buy-pass.ina
# ilay  boky. 
# DX.sg a book(s).
# Lit. "Was said by Paul that will be bought the (previous mention)
# book." "Paul said that (he = Paul) would buy the (previous mention) book."
# This is the preferred interpretation in the absence of an elaborate
# discourse context with a verb of saying.
# Note that the complementizer fa is optional.
# Note the absence of the clitic pronoun -ny 'by-him/her' on the lower
# passive voice verb.
# Note that there is no coreferential pronoun accompanying the lower verb and
# that [  hovidina ilay boky ] represents a sentential subject of the matrix
# passive verb nolazaina.


# The case of Control structures and empty subjects:
# In Malagasy there is a clear-cut distinction between a Control structure and a Raising-
# to-Subject construction;

  Mikasa hanasa ny vilia Rakoto.
# M-i-kasa              h-an-(s)asa           ny         vilia  ø     Rakoto.
# Pres-pref.i-root.kasa fut-pref.an-root.sasa Det.unspec dishes empty Rakoto
# 'Rakoto intends to wash the dishes' (Original translation).
# This sequence has been adapted from M. Pearson (2001a: 93, ex.23); where there are two
# different verbs and where the lower predicate 'hanasa''fut-wash' does not have an
# overt grammatical subject, the NP 'ny vilia''the dishes' being a direct object;
# however note that this is a slightly different empty subject position and that
# is why a different symbol, i.e. ø is used instead of the usual [ ø ] above. For
# further detail consult Randriamasimanana (1994: 488).

  Kasain-dRakoto hosasana ny vilia.
# Kasain-dRakoto             [ hosasana             ] ny         vilia.
# be.intended-by.ep.d-Rakoto [ passive.fut.be.washed] Det.unspec dishes
# 'The dishes, Rakoto intends to wash' (Original translation).
# This sequence has been adopted from M. Pearson (2001: 93, ex.23); where there are two
# different verbs; now M. Pearson analyses the sentence as indicated by the square
# brackets, i.e. the derived subject 'ny vilia''the dishes' has been raised, lifted
# out of the embedded clause comprising the lower verb 'hosasana' into the higher,
# matrix verbal clause with 'kasaina';
# this sequence is acceptable in Colloquial Malagasy;

  Nikasa hanasa ny vilia Ingahy Rakoto.
# N-i-kasa              [ h-an-(s)asa           ny         vilia  ø ]     Ingahy Rakoto.
# Past-pref.i-root.kasa [ fut-pref.an-root.sasa Det.unspec dishes empty ] D.sg   Rakoto
# 'Mr. Rakoto intended to wash the dishes'.
# Note that the lower verb hanasa 'will-wash' is not accompanied by an overt subject,
# the NP 'Ingahy Rakoto' being the subject of the higher, matrix verb nisaka 'intended';
# the relevant structure is indicated by the square brackets with an empty subject in
# the lower clause;

  Nokasain' Ingahy Rabe hosasana ny vilia.
# No-kasa-in(a) Ingahy Rabe                        [ ho-sasana             ny         vilia ].
# Passive.past-be.intended-suffix.ina-ep.d-by.D.Rabe passive.fut-be.washed Det.unspec dishes
# Literally 'The dishes will be washed was intended by Mr.Rabe', i.e. 'Mr.Rabe intended to wash
# the dishes';
# there is no raising, lifting of the subject of the embedded subject 'ny vilia' into the
# matrix clause; for further illustration of this type of analysis consult Randriamasimanana
# (2002: 57-58, 65-66);
# this is the standard Malagasy equivalent to a similar sequence shown earlier;

  Mety hanasa vilia Ingahy Rakoto.
# Mety   [ h-anasa  vilia  ø  ]   Ingahy Rakoto.
# likely [ fut-wash dishes empty] D.sg   Rakoto
# Lit. 'It is likely that Mr. Rakoto will wash dishes', i.e.'Mr. Rakoto is likely to wash dishes.'
# Mety is a raising-to-subject predicate as described in Randriamasimanana (1986: 539-553);
# this is a different type of construction than the one involving subject Control 'mikasa''intend';
# for further detail consult Randriamasimanana (1994: 488, also 493, ex. (25)a & (25)b.).
# It is worth pointing out that in this last sequence, the nature of the empty position is totally
# different from the previous two cases.




# Clitic forms attached to a lexical head, i.e. a V, an N (or NP) or a Prep.
# Distribution of independent pronouns and clitic forms: In general an independent pronoun
# form can show up as a grammatical subject or a direct object; by contrast, a clitic
# has to be attached to a passive head V or as a genitive case suffixed to either an N or
# to the general preposition 'tamin(a)', which has a wide range of meanings depending on
# the construction in which it shows up, i.e. on, under, to the side, below, with, etc...;
# as for the zero pronoun form, it usually shows up as a grammatical subject whether the
# verb is in the active or passive voice;

# The distribution of independent pronoun forms  as well as zero pronoun was already 
# illustrated above in Part A; in Part C, we will illustrate the case of clitics;
# with regard to clitics --as is the case with epenthetic 'na'-- a clitic does not require 
# Stress Shift, expected of suffix 'ina' as  explained in the on-line document # Malagasy 
# phonetics/phonology & morphology: Stress assignment: ESRC-OX-05-CR109. In addition, one 
# other on-line document labelled # Malagasy phonetics/phonology & morphology: Malagasy ka,
# tra, na endings of words: ESRC-OX-05-CR104 does provide numerous illustrative examples 
# involving epenthetic 'na'.


# Clitic forms attached to passive Vs:
# The relevant clitic forms to consider are: 

# for simple/unmarked forms, -ko 'by.me', -nay 'by.us.exclusive', -ntsika 'by.us.inclusive',
#  -nao 'by.you.sing' -nareo 'by.you.plural',  -ny 'by.him/her.or.by.them.human';

# for complex/marked forms involving genitive-marker n(a) followed by izy + appositive N, izy ireo, etc...
# when the [ + Animate ] & [ + Plural ] features are involved; otherwise, when the relevant feature is
# [ - Animate ], the appropriate form is 'ilay izy' following the genitive marker n(a) or in the case of
# a plural, the form is ireo following the n(a)-genitive marker;


  Nilaza  Rabe fa novonoiko tamin'ity antsy ity ilay akoho.
# N-i-laza     Rabe fa   no-       vono-i(na)-   ko        
# past-prf-say Rabe comp past.pass-kill-pass.ina-by.me
# t-amin'   ity  antsy  ity   ilay akoho.
# perf-with DX-sg knife DX-sg DX-sg chicken.
# Lit. "Rabe said that was killed by me with this knife the (previous mention)
# chicken.""Rabe said that I the speaker killed the (previous mention) chicken
# with this knife." Note the first person singular pronoun clitic -ko 'of.me' 
# attached to the passive verb 'novonoiko','no + vono + i(na) + ko'
# where the morpheme -i- represents the passive 1 suffix form 'ina'.


  Fantatro fa kinga i Paoly.
# Fantatr-      o     fa   kinga i    Paoly.
# root.be.known.to.me comp smart D.sg Paul
# Lit. "Is known to me that Paul is smart." "I know that Paul is smart."
# Note that 'fantatra' 'be-known' is a root verb, which usually does not
# take any affix; in this instance, the genitive first person singular pronoun
# clitic -ko 'of.me' is reduced to the suffix vowel 'o' since the verb ends in
# the syllable 'tra', one of three which require a special phonological treatment.
# Fantatr(a) + (k)o > fantatro, where suffix 'o' represents a first person singular
# clitic pronoun.


  Nolazain'i Paoly fa hovidiny ilay boky.
# No-       laza-in(a)       i    Paoly fa   ho-      vid(i)-i(na)-   ny         
# past.pass-say- pass.ina.by D.sg Paul  comp fut.pass-buy-   pass.ina-by.him/her
# ilay  boky. 
# DX-sg a book(s).
# Lit. "Was said by Paul that will be bought by him the (previous mention)
# book." "Paul said that he (= Paul) would buy the (previous mention) book."
# This is the preferred interpretation in the absence of an elaborate
# discourse context. Note the genitive form of the third person singular pronoun
# -ny cliticised to the embedded verb 'hovidiny''future.passive.ho + root.vid(i)
# + passive.suffix.ina'.
# Note the clitic pronoun -ny 'by-him/her' attached to the lower passive voice verb,
# which refers back to the genitive NP -'i Paoly'.


  Nokapohin'i Jaona ny zaza maditra.
# No-kapohi-in(a)                                      i    Jaona ny         zaza       maditra.
# passive.past.PUNCTUAL-stem.kapohi-passive.suffix.ina D.sg John  Det.unspec child(ren) unruly
# '(All) unruly children were deliberately punished by John.'
# The root is kapoka N 'strike, beating'; the passive 1 imperative is kapohy '(let-s.one/sth)-
# be-struck-by.you'; passive 1 indicative form kapohina kapoh(i)+ina 'stem.kapohi-suffix.ina';

  Nokapohiny ny zaza maditra.
# No-kapoh(i)-i-ny                                             i    Jaona ny         zaza       maditra.
# passive.past.PUNCTUAL-stem.kapohi-passive.suffix.i-clitic.ny D.sg John  Det.unspec child(ren) unruly
# '(All) unruly children were deliberately punished by him/her.'
# Note that the suffix.i represents the passive voice whereas it is the ending na of passive.ina, which has 
# been replaced with the pronoun clitic -ny 'by.him/her';

  Nokapohin'izy mivady ny zaza maditra.
# No-kapohi-in(a)                                      izy mivady   ny         zaza       maditra.
# passive.past.PUNCTUAL-stem.kapohi-passive.suffix.ina they spouses Det.unspec child(ren) unruly
# '(All) unruly children were deliberately punished by the two spouses.'
# Note the complex form izy mivady, i.e. izy + mivady 's/he + spouses';

  Matin'ny varatra ilay olona.
# Mati-n(a)   ny varatra ilay    olona.
# dead-due.to Det.unspec thunder DX.sg person
# 'The (previous mention) person is (accidentally) dead because of thunder.'
# Note the epenthetic na 'due.to' attached to the root mati 'dead'.

  Matin' ilay izy ilay olona.
# Mati-n(a)   ilay izy ilay  olona.
# dead-due.to proDX    DX.sg person
# 'The (previous mention) person is (accidentally) dead because of it (nonanimate).'
# Note the complex form 'ilay izy' to refer to a nonanimate entity;

  Matiny ilay olona.
# Mat(i)-i(na)-ny                  ilay  olona.
# dead-due.to-passive.i-clitic.ny  DX.sg person
# 'The (previous mention) person is (possibly accidentally) dead because of him/her (nanimate).'
# This sequence cannot be interpreted to refer to a non-animate referent for the clitic -ny;
# the clitic -ny by default refers to an animate entity;


# Clitics attached to Ns or NPs:

  Hevitr'i Jeanne
# Hevitr'i Jeanne
# idea(s)-of D.sg Jeanne
# 'Jeanne's ideas';

  Hevitra kitoatoan'i Jeanne
# Hevitra kitoatoa-n(a)    i Jeanne
# idea(s) adj.quick.fix-of D.sg Jeanne
# 'Jeanne's quickfix ideas'
# Note the epenthetic na attached to the right of NP, not N.


# Clitics attached to Prepositions:

  Nody tany aminy i Jeanne.
# N-ody         t-any      ami-ny                       i    Jeanne.
# past-root.ody past-there Obl.case.ami(na)-clitic.ny D.sg Jeanne
# Literally 'Jeanne returned gone to.place-of-hers', i.e. 'Jeanne went home.'
# Note that the clitic -ny in the above sequence can and does refer to 'i Jeanne';
# such an interpretation rests on the fact that the PP 'tany aminy' is an adjunct to
# the first verb 'nody''went.back'; note that t- on Prep tany is a past tense-marker;
# we have the configuration past.tense ... past.tense, the hallmark of adjunction in
# in Malagasy.
# In addition note that the preposition 'aminy' is now re-interpreted as an Oblique
# case-marker and as a result no longer carries the morpheme t-. See on-line document
# relative to Malagasy prepositions.

  Tany aminy i Jeanne.
# T-any      ami-ny                       i    Jeanne.
# Perf-there Obl.case.prep.ami(na)-clitic.ny D.sg Jeanne
# Literally 'Jeanne went away gone to.place-of-hers', i.e. 'Jeanne went to somebody else's home.'
# Note that the clitic -ny in this other sequence cannot and does not refer to 'i Jeanne';
# the morpheme t- on Prep tany is ambiguous between a perfective aspect-marker, as is obligatory
# in the embedded structure shown immediately below and a past tense-marker, as is possible when
# we have an independent clause like the one under consideration above; with t- as a past tense-
# marker, tany literally means 'went to...', a Motion verb.
# Again note that the preposition 'aminy' is now re-interpreted as an Oblique
# case-marker and as a result no longer carries the morpheme t-.See on-line document
# relative to Malagasy prepositions.

  Nandeha tany aminy i Jeanne.
# N-an-(l>d)eha          t-any      ami-ny                       i    Jeanne.
# past-pref.an-root.leha perf-there Obl.case.prep.ami(na)-clitic.ny D.sg Jeanne
# Literally 'Jeanne went gone to.place-of-someone.else', i.e. 'Jeanne went to his/her home.'
# The clitic -ny is obligatorily disjoint in reference from 'i Jeanne';
# note that the morpheme t- on Prep tany is an aspectual marker; therefore the PP tany aminy
# is an argument of the higher verb nandeha 'went'; under these conditions it is normal that
# the clitic pronoun -ny canNOT refer to the subject NP 'i Jeanne'.
# Again note that the preposition 'aminy' is now re-interpreted as an Oblique
# case-marker and as a result no longer carries the morpheme t-.See on-line document
# relative to Malagasy prepositions.


# Notes relative to putative genitive antecedents of reflexives in Malagasy.

# Matt Pearson in his 2001 dissertation, The clause structure of Malagasy: A minimalist approach. UCLA 
# Dissertations in Linguistics 21 produces the following sequences:

# Ungrammatical:
  Novonoin'ny lehilahy ny tenany. (0! 0 0 0)
# No-vono-in(a)                   ny         lehilahy ny         tena-ny
# NonPerf.Past-root.vono-pass.ina Det.unspec men      Det.unspec humble.self-his/hers
#  'The man killed himself.'
# The above sequence is adapted from Chapter 3, example (14) and is at least of very dubious
# grammaticality, given the non-perfective aspect of the verb 'novonoina' inside the clause.
# The genitive 'ny lehilahy' is the putative antecedent for the anaphor 'ny tenany''self'.
# Note that the reflexive 'ny tenany''self' occupies the grammatical subject position,
# that the verb root vono 'killing' is accompanied by the non-perfective aspect passive voice 1
# form no ...ina and that the grammatical subject has not been fronted.

# Ungrammatical:
  Namonoan'ny lehilahy tena ny zanany.(0! 0 0 0)
# N-a(n>m)-(v)ono-an(a)             ny         lehilahy tena ny         zana(ka)-ny.
# Past-pref.an-root.vono-suffix.ana Det.unspec men      self Det.unspec children
#'The man killed himself for his children.'
# The above sequence is from Chapter 3, example (38). This is irretrievably ungram-
# matical and sounds pretty much like Foreigner Talk.
# The genitive 'ny lehilahy' is the putative antecedent for the anaphor 'tena''self'.
# Note that the reflexive 'tena''self' occupies the grammatical object position and
# that the verb root vono 'killing' is accompanied by the non-perfective aspect passive voice 2
# form an ...ana.



# Typical use of 'ny tenako''my humble self' from Rajemisa-Raolison (1995: 938):

  Tsy naniraka olona izy fa ny tenany mihitsy no tonga.
# Tsy n-an-iraka        olona  izy  fa  ny  tena-ny            mihitsy no    tonga.
# Neg past-pref.an-send people s/he but the humble.self-of.his indeed  focus perf.arrived
# 'S/he did not send (other) people, but s/he (him/herself) came in person.'
# Note that 'ny tenany' occupies the grammatical subject position and is used as a polite
# way of referring to 'I''first.person.singular'. Also note that the subject has been
# fronted inside the embedded clause and finally that the embedded verb is in the perfective
# aspect, i.e. tonga is a passive root with an inherent perfective aspect feature.
# There is no overt antecedent to the reflexive form 'ny tenany'.

# At least, two of the above features were already noted in Randriamasimanana (1986: 232, ex (1)):
  Tonga ny tenako.
# Tonga   ny  tena-ko.
# Arrived the body-of-me
# 'I (my humble person) have arrived.'
# Note the perfective aspect verb 'tonga''has.arrived' and that there is no overt antecedent to
# the reflexive form 'ny tenako'.


# Typical uses of 'ny tenako''my humble self/person/body' from Malzac (1888: 694):
  Ny tenany ihany no tao.
# Ny  tena-ny          ihany no   t-ao.
# The body-of-his/hers only focus perf-there
# (Original French translation)'Il est venu là en personne'. (English)'He was there by himself.'
# Note the perfective aspect-marker t- on the preposition 'tao'; also the grammatical subject
# has been fronted.

  Mararirary be ihany ny tenako rehetra.
# Mararirary    be          ihany ny  tena-ko      rehetra.
# Somewhat.sick pretty.much only  the body-of-mine whole
# (Original French translation)'Je souffre dans tout mon corps.' (English)'My whole body is
# pretty sore.' Note the literal translation of ny tenako 'the body-of-mine'.





# Charles Randriamasimanana, PhD in Linguistics (Malagasy, Austronesian)